ChatGPT in Scientific Research and Writing: A Beginner’s Guide

The developers of ChatGPT have predicted that, within the next ten years, artificial intelligence (AI) systems will exceed expert skill levels in most domains, and carry out as much productive work as one of today’s largest corporations. Since the public release of ChatGPT, there has been surging interest in exploring the use of large language models, including ChatGPT, in scientific research, publication, and science communication in general. In this book, we will explore the models’ capabilities, including GPT-4, GPT-3.5, and GPT-enabled new Bing (now Copilot), for carrying out the tasks through different stages of scientific research from research conceptualization, study design, to publication and science communication. We used these models for abstracting key points and extracting specific information from research publications, interpreting figures, evaluating research papers, spotting errors, responding to reviewer’s comments, language editing, designing experiments, creating survey questionnaires, brainstorming, writing research proposals, and creating visuals. Major limitations of these models include hallucinations, randomness in answers when prompted by identical questions, and the lack of support for big data scrapping, processing, and visualization.